aliasmethod 0.4.0

Implementation of Walker's Alias method.
Documentation 
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
86
87
88
89
90
91
92
93
94
95

extern crate aliasmethod;
extern crate rand;

#[cfg(test)]
mod tests {
    use aliasmethod::AliasTable;

    use rand::thread_rng;
    use std::collections::HashMap;

    struct TestParam {
        weights: Vec<f64>,
        rates: Vec<f64>,
    }

    #[test]
    fn test_probability() {
        let params: [TestParam; 6] = [
            TestParam {
                weights: vec![10.0, 15.0],
                rates: vec![40.0, 60.0],
            },
            TestParam {
                weights: vec![20.0, 30.0],
                rates: vec![40.0, 60.0],
            },
            TestParam {
                weights: vec![20.0, 5.0],
                rates: vec![80.0, 20.0],
            },
            TestParam {
                weights: vec![25.0],
                rates: vec![100.0],
            },
            TestParam {
                weights: vec![1.0, 99.0],
                rates: vec![1.0, 99.0],
            },
            TestParam {
                weights: vec![1.0, 1.0, 8.0],
                rates: vec![10.0, 10.0, 80.0],
            },
        ];

        for (
            param_no,
            &TestParam {
                ref weights,
                ref rates,
            },
        ) in params.into_iter().enumerate()
        {
            match AliasTable::new(&weights) {
                Err(e) => {
                    assert!(false, "error : {}", e);
                }
                Ok(alias_table) => {
                    let sample: i64 = 100000;

                    let mut results: HashMap<usize, i64> = HashMap::new();

                    for _ in 0..sample {
                        let n = alias_table.random(thread_rng());
                        let count = results.entry(n).or_insert(0);
                        *count += 1;
                    }

                    let fs = sample as f64;

                    for (n, rate) in rates.into_iter().enumerate() {
                        match results.get(&n) {
                            None => {
                                assert!(false, "Not found '{}' in results", n);
                            }
                            Some(count) => {
                                let p: f64 = rate / 100.0;
                                let q: f64 = 1.0 - p;

                                let expected = fs * p;
                                // 3.89 = inverse of normal distribution function with alpha=0.9999
                                let delta = 3.89 * (fs * p * q).sqrt();

                                let count = *count as f64;
                                let ok = expected - delta <= count && count <= expected + delta;
                                assert!(ok, format!(
                                        "The probability is out of by interval estimation. param_no={} num={} weight={} actual={}, expected={}, delta={}",
                                        param_no, n, weights[n as usize], count, expected, delta));
                            }
                        }
                    }
                }
            }
        }
    }
}